文献类型：期刊文献

中文题名：图书馆在架图书的索书号图像提取与分割

英文题名：Extraction and Segmentation of Call Number Image for Books on Shelves

作者：方建军[1];赵强强[2]

第一作者：方建军

机构：[1]北京联合大学自动化学院;[2]北方工业大学机电工程学院

第一机构：北京联合大学城市轨道交通与物流学院

年份：2015

卷号：29

期号：1

起止页码：87-92

中文期刊名：北京联合大学学报

基金：北京市属市管高校学术创新团队项目(PHR201107149)

语种：中文

中文关键词：图书馆机器人;索书号;概率Hough变换;颜色特征

外文关键词：Library robot; Book call number; Probabilistic Hough transform; Color features

摘要：图书索书号提取与分割是图书馆自动存取机器人完成图书上下架的关键技术。提出了一种利用索书号颜色特征和书脊边缘信息进行图像提取分割的方法,有效地解决了背景相近的相邻索书号之间的分割问题。通过实验利用HSV颜色空间提取索书号区域,用Canny算子做书脊边缘检测,采用概率霍夫变换法查找书脊边缘线段,实现了索书号图像的提取与分割,并针对厚、较厚、中等厚、较薄、薄、混合等几种类型的书脊进行了分类实验和研究,可以有效地提取出不同情况下每本图书的索书号,为图书馆自动存取机器人完成自动存取操作奠定了基础。
The extraction and segmentation of book call number is a key technology for book retrieval and return robot putting up and down the books on the shelves. This paper presents an image extraction and segmentation method using color feature and edge information, effectively solving the segmentation problem of similar background between adjacent book call number. Through the HSV color space extracting book call area, using Canny to do the spine edge detection, using probabilistic Hough transform method to find the line of spine edge, the extraction and segmentation of book call number is realized. And through the classified experiments on the thick, medium thick, thin, little thin, hybrid and other types of spine, which are effectively extracted in different circumstances, the method lays the foundation for book retrieval and return robot completing the automatic access operation.